AgriVision
AI-powered precision agriculture with satellite analytics
Overview
AgriVision is a full-stack precision-agriculture platform combining Sentinel-2 satellite imagery, IoT sensor telemetry, and AI to give farmers concrete, localized insights. The dashboard is a React/Vite app styled with Tailwind and wired up with Recharts and Leaflet for satellite analytics; the backend is an Express API with controllers and services for NDVI/analytics ingestion, AI descriptions, and PDF report generation. It integrates Firebase Auth + Realtime DB for identity and sync, Sentinel Hub for imagery, and Gemini/Groq for LLM features — including the Kisan Mitra assistant. The UI is multilingual (English, Hindi, Bengali) and designed for rural connectivity. AgriVision was selected as a national finalist at the Smart India Hackathon.
Architecture
- 01Frontend: React/Vite dashboard with Tailwind, Recharts, Leaflet.
- 02Backend: Express controllers/services for reports, AI descriptions, and NDVI ingestion.
- 03Firebase Auth + Realtime DB; Sentinel Hub for imagery; Gemini/Groq for AI.
What I learned
- Working with geospatial data and vegetation-index math.
- Designing a multilingual UX around noisy rural connectivity.
- Integrating multiple LLM providers behind a stable service interface.